tf.keras.metrics.SparseCategoricalCrossentropy
Stay organized with collections
Save and categorize content based on your preferences.
Computes the crossentropy metric between the labels and predictions.
tf.keras.metrics.SparseCategoricalCrossentropy(
name='sparse_categorical_crossentropy', dtype=None, from_logits=False, axis=-1
)
Use this crossentropy metric when there are two or more label classes.
We expect labels to be provided as integers. If you want to provide labels
using one-hot
representation, please use CategoricalCrossentropy
metric.
There should be # classes
floating point values per feature for y_pred
and a single floating point value per feature for y_true
.
In the snippet below, there is a single floating point value per example for
y_true
and # classes
floating pointing values per example for y_pred
.
The shape of y_true
is [batch_size]
and the shape of y_pred
is
[batch_size, num_classes]
.
Usage:
# y_true = one_hot(y_true) = [[0, 1, 0], [0, 0, 1]]
# logits = log(y_pred)
# softmax = exp(logits) / sum(exp(logits), axis=-1)
# softmax = [[0.05, 0.95, EPSILON], [0.1, 0.8, 0.1]]
# xent = -sum(y * log(softmax), 1)
# log(softmax) = [[-2.9957, -0.0513, -16.1181],
# [-2.3026, -0.2231, -2.3026]]
# y_true * log(softmax) = [[0, -0.0513, 0], [0, 0, -2.3026]]
# xent = [0.0513, 2.3026]
# Reduced xent = (0.0513 + 2.3026) / 2
m = tf.keras.metrics.SparseCategoricalCrossentropy()
_ = m.update_state([1, 2],
[[0.05, 0.95, 0], [0.1, 0.8, 0.1]])
m.result().numpy()
1.1769392
m.reset_states()
_ = m.update_state([1, 2],
[[0.05, 0.95, 0], [0.1, 0.8, 0.1]],
sample_weight=tf.constant([0.3, 0.7]))
m.result().numpy()
1.6271976
Usage with tf.keras API:
model = tf.keras.Model(inputs, outputs)
model.compile(
'sgd',
loss='mse',
metrics=[tf.keras.metrics.SparseCategoricalCrossentropy()])
Args |
name
|
(Optional) string name of the metric instance.
|
dtype
|
(Optional) data type of the metric result.
|
from_logits
|
(Optional ) Whether y_pred is expected to be a logits tensor.
By default, we assume that y_pred encodes a probability distribution.
|
axis
|
(Optional) Defaults to -1. The dimension along which the metric is
computed.
|
Args |
fn
|
The metric function to wrap, with signature
fn(y_true, y_pred, **kwargs) .
|
name
|
(Optional) string name of the metric instance.
|
dtype
|
(Optional) data type of the metric result.
|
**kwargs
|
The keyword arguments that are passed on to fn .
|
Methods
reset_states
View source
reset_states()
Resets all of the metric state variables.
This function is called between epochs/steps,
when a metric is evaluated during training.
result
View source
result()
Computes and returns the metric value tensor.
Result computation is an idempotent operation that simply calculates the
metric value using the state variables.
update_state
View source
update_state(
y_true, y_pred, sample_weight=None
)
Accumulates metric statistics.
y_true
and y_pred
should have the same shape.
Args |
y_true
|
Ground truth values. shape = [batch_size, d0, .. dN] .
|
y_pred
|
The predicted values. shape = [batch_size, d0, .. dN] .
|
sample_weight
|
Optional sample_weight acts as a
coefficient for the metric. If a scalar is provided, then the metric is
simply scaled by the given value. If sample_weight is a tensor of size
[batch_size] , then the metric for each sample of the batch is rescaled
by the corresponding element in the sample_weight vector. If the shape
of sample_weight is [batch_size, d0, .. dN-1] (or can be broadcasted
to this shape), then each metric element of y_pred is scaled by the
corresponding value of sample_weight . (Note on dN-1 : all metric
functions reduce by 1 dimension, usually the last axis (-1)).
|
Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.
Last updated 2020-10-01 UTC.
[[["Easy to understand","easyToUnderstand","thumb-up"],["Solved my problem","solvedMyProblem","thumb-up"],["Other","otherUp","thumb-up"]],[["Missing the information I need","missingTheInformationINeed","thumb-down"],["Too complicated / too many steps","tooComplicatedTooManySteps","thumb-down"],["Out of date","outOfDate","thumb-down"],["Samples / code issue","samplesCodeIssue","thumb-down"],["Other","otherDown","thumb-down"]],["Last updated 2020-10-01 UTC."],[],[],null,["# tf.keras.metrics.SparseCategoricalCrossentropy\n\n\u003cbr /\u003e\n\n|--------------------------------------------------------------------------------------------------------|------------------------------------------------------------------------------------------------------------------------------|\n| [TensorFlow 1 version](/versions/r1.15/api_docs/python/tf/keras/metrics/SparseCategoricalCrossentropy) | [View source on GitHub](https://github.com/tensorflow/tensorflow/blob/v2.2.0/tensorflow/python/keras/metrics.py#L3048-L3117) |\n\nComputes the crossentropy metric between the labels and predictions.\n\n#### View aliases\n\n\n**Main aliases**\n\n[`tf.metrics.SparseCategoricalCrossentropy`](/api_docs/python/tf/keras/metrics/SparseCategoricalCrossentropy)\n**Compat aliases for migration**\n\nSee\n[Migration guide](https://www.tensorflow.org/guide/migrate) for\nmore details.\n\n[`tf.compat.v1.keras.metrics.SparseCategoricalCrossentropy`](/api_docs/python/tf/keras/metrics/SparseCategoricalCrossentropy)\n\n\u003cbr /\u003e\n\n tf.keras.metrics.SparseCategoricalCrossentropy(\n name='sparse_categorical_crossentropy', dtype=None, from_logits=False, axis=-1\n )\n\nUse this crossentropy metric when there are two or more label classes.\nWe expect labels to be provided as integers. If you want to provide labels\nusing `one-hot` representation, please use `CategoricalCrossentropy` metric.\nThere should be `# classes` floating point values per feature for `y_pred`\nand a single floating point value per feature for `y_true`.\n\nIn the snippet below, there is a single floating point value per example for\n`y_true` and `# classes` floating pointing values per example for `y_pred`.\nThe shape of `y_true` is `[batch_size]` and the shape of `y_pred` is\n`[batch_size, num_classes]`.\n\n#### Usage:\n\n # y_true = one_hot(y_true) = [[0, 1, 0], [0, 0, 1]]\n # logits = log(y_pred)\n # softmax = exp(logits) / sum(exp(logits), axis=-1)\n # softmax = [[0.05, 0.95, EPSILON], [0.1, 0.8, 0.1]]\n # xent = -sum(y * log(softmax), 1)\n # log(softmax) = [[-2.9957, -0.0513, -16.1181],\n # [-2.3026, -0.2231, -2.3026]]\n # y_true * log(softmax) = [[0, -0.0513, 0], [0, 0, -2.3026]]\n # xent = [0.0513, 2.3026]\n # Reduced xent = (0.0513 + 2.3026) / 2\n m = tf.keras.metrics.SparseCategoricalCrossentropy()\n _ = m.update_state([1, 2],\n [[0.05, 0.95, 0], [0.1, 0.8, 0.1]])\n m.result().numpy()\n 1.1769392\n\n m.reset_states()\n _ = m.update_state([1, 2],\n [[0.05, 0.95, 0], [0.1, 0.8, 0.1]],\n sample_weight=tf.constant([0.3, 0.7]))\n m.result().numpy()\n 1.6271976\n\nUsage with tf.keras API: \n\n model = tf.keras.Model(inputs, outputs)\n model.compile(\n 'sgd',\n loss='mse',\n metrics=[tf.keras.metrics.SparseCategoricalCrossentropy()])\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n| Args ---- ||\n|---------------|-----------------------------------------------------------------------------------------------------------------------------------------|\n| `name` | (Optional) string name of the metric instance. |\n| `dtype` | (Optional) data type of the metric result. |\n| `from_logits` | (Optional ) Whether `y_pred` is expected to be a logits tensor. By default, we assume that `y_pred` encodes a probability distribution. |\n| `axis` | (Optional) Defaults to -1. The dimension along which the metric is computed. |\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n| Args ---- ||\n|------------|-----------------------------------------------------------------------------|\n| `fn` | The metric function to wrap, with signature `fn(y_true, y_pred, **kwargs)`. |\n| `name` | (Optional) string name of the metric instance. |\n| `dtype` | (Optional) data type of the metric result. |\n| `**kwargs` | The keyword arguments that are passed on to `fn`. |\n\n\u003cbr /\u003e\n\nMethods\n-------\n\n### `reset_states`\n\n[View source](https://github.com/tensorflow/tensorflow/blob/v2.2.0/tensorflow/python/keras/metrics.py#L218-L224) \n\n reset_states()\n\nResets all of the metric state variables.\n\nThis function is called between epochs/steps,\nwhen a metric is evaluated during training.\n\n### `result`\n\n[View source](https://github.com/tensorflow/tensorflow/blob/v2.2.0/tensorflow/python/keras/metrics.py#L376-L386) \n\n result()\n\nComputes and returns the metric value tensor.\n\nResult computation is an idempotent operation that simply calculates the\nmetric value using the state variables.\n\n### `update_state`\n\n[View source](https://github.com/tensorflow/tensorflow/blob/v2.2.0/tensorflow/python/keras/metrics.py#L574-L605) \n\n update_state(\n y_true, y_pred, sample_weight=None\n )\n\nAccumulates metric statistics.\n\n`y_true` and `y_pred` should have the same shape.\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n| Args ||\n|-----------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|\n| `y_true` | Ground truth values. shape = `[batch_size, d0, .. dN]`. |\n| `y_pred` | The predicted values. shape = `[batch_size, d0, .. dN]`. |\n| `sample_weight` | Optional `sample_weight` acts as a coefficient for the metric. If a scalar is provided, then the metric is simply scaled by the given value. If `sample_weight` is a tensor of size `[batch_size]`, then the metric for each sample of the batch is rescaled by the corresponding element in the `sample_weight` vector. If the shape of `sample_weight` is `[batch_size, d0, .. dN-1]` (or can be broadcasted to this shape), then each metric element of `y_pred` is scaled by the corresponding value of `sample_weight`. (Note on `dN-1`: all metric functions reduce by 1 dimension, usually the last axis (-1)). |\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n| Returns ||\n|---|---|\n| Update op. ||\n\n\u003cbr /\u003e"]]